From Meta, a crackdown on scams that use AI to impersonate brands and celebrities.

Meta has announced new countermeasures to combat fraud on its platforms that use artificial intelligence.

The tools are designed to identify in real time the identity theft of brands and celebrities, blocking deceptive links and so-called "celeb bait" content, videos created with AI to deceive the public.

The plan includes the introduction of targeted security alerts: Facebook will flag suspicious friend requests, Messenger will monitor anomalous accounts, and WhatsApp will send notifications for every attempt to connect new devices.

"We have developed advanced artificial intelligence systems capable of analyzing multiple signals," reads an official statement, "such as text, images, and surrounding context, to identify a wider range of sophisticated scam patterns, more quickly and at scale."

At the same time, the company intends to tighten admission criteria for advertisers, aiming to derive 90% of advertising revenue from verified partners by the end of 2026.

According to an online report by Reuters released last November, Meta is said to have earned billions of dollars each year from ads that allegedly promoted scams and illegal products on its platform.

In response, the giant stated that it has reduced reports of these ads globally by 58% in 2025.

The offensive from Menlo Park follows a series of legal actions against international entities engaged in financial fraud.

Data confirms the magnitude of the phenomenon: Meta removed 159 million fraudulent ads last year and deactivated about 11 million accounts linked to criminal organizations.